-
Notifications
You must be signed in to change notification settings - Fork 4
Hode vs. property graphs
Hode lets you represent information no other editor permits. A property graph lets you attach properties to an object, and lets you pair objects in relationships. It does not let you represent relationships between relationships.
Bob #watches NBC (where #watches is the relationship), and age(Bob) = 63 (where age is a property`) are representable in a property graph.
But Bob #buys apples #from Costco is not representable -- it would require an edge to connect more than two vertices. Similarly, a second-order relationship like Bob #buys diapers ##because Bob #(has a) baby (where ##because is the second-order relationship between the two first-order relationships) is not representable -- it would require an edge to connect two edges.
Hode lets you represent your information using a language, Hash, mostly of your own invention -- you only need to learn how to use the # symbol to represent joints in relationships. Hash can closely resemble ordinary natural language. Querying your data is almost as simple.
Hode also lets you traverse the data "edgewise", from relationships to their member expressions, and from expressions to the relationships containing them. It is an editor, after all, not a programming language. I believe it is much easier to use than something like Neo4j or Titan.
I don't know how well it scales. Surely not as well as an established graph database. If you're using it for a personal knowledge base, you're probably fine; if you're using it to compete with Google, it's not there yet.
It's still alpha. It will evolve, and to my knowledge so far it has only one user, although a few more people say they're excited to try it.